<PRE>
<PAGE>   1

                       Filed by Marvell Technology Group Ltd.
                       Filed  pursuant to Rule 425 under the  Securities  Act of
                       1933 and Rule 14a-12  under the  Securities  Exchange Act
                       of 1934
                       Subject Company:  Galileo Technology Ltd.
                       Commission File No.: 0-30877

The following is the script from a joint conference call conducted by Marvell
Technology Group Ltd., and Galileo Technology Ltd. on October 17, 2000.

================================================================================


                            MARVELL TECHNOLOGY GROUP
                        MARVELL / GALILEO CONFERENCE CALL

                                October 17, 2000
                                 10:00 a.m. EDT


CHAIRPERSON: Sehat Sutardja, Co-Chairman, President and CEO

<TABLE>
<S>                        <C>                      <C>
[OP = Operator             SS = Sehat Sutardja      GH = George Hervey
 AW = Avigdor Willenz      MS = Mike Seth           SP = Unidentified Speaker]
</TABLE>

OP: Ladies and gentlemen, thank you for standing by. Welcome to the Marvell
Technology / Galileo Technology Conference Call. At this time, all participants
are in a listen-only mode. Later we will conduct a question and answer session.
At that time, if you have a question, you will need to press the "1" followed by
the "4" on your telephone. As a reminder, this conference is being recorded
today, Tuesday, October 17, 2000. I would now like to turn the conference over
to George Hervey, Chief Financial Officer of Marvell Technology. Please go
ahead, sir.

GEORGE HERVEY

     Thank you. Good morning, this is George Hervey, Chief Financial Officer of
Marvell. Thank-you for joining us today for the Marvell Conference call to
discuss the acquisition of Galileo. In addition, we will discuss Galileo's third
quarter results.

     You should have received a press release via fax. It is also available at
our web site at www.marvell.com, and over BusinessWire.

     The participants on the call with me today are Dr. Sehat Sutardja,
Co-Chairman, President, and Chief Executive Officer of Marvell; Avigdor Willenz,
Chairman and Chief Executive Officer of Galileo. Sehat will first give a brief
overview of today's announcement and the immediate opportunities that we see for
the combined company. Next, Avigdor will discuss Galileo's decision to join the
Marvell team, and thoughts on longer-term synergies created by the combination.

     We will offer some brief remarks to summarize why we're so excited about
this transaction, then we will follow with a discussion of Galileo's third
quarter results. Immediately following each discussion, we will open the floor
for Q&A. First we will take questions about the merger, then any questions you
may have about Galileo's earnings results.

     Before we get started, we would appreciate your attention as we go through
the following. The first portion of the conference call may contain
forward-looking statements within the meaning of the Safe Harbor provisions of
the Private Securities Litigation Reform Act of 1995. These statements are based
on the current expectations or beliefs of Marvell and Galileo's management, and
are subject to a number of factors and uncertainties that could cause

<PAGE>   2

MARVELL TECHNOLOGY GROUP -- October 17, 2000                              Page 2


actual results to differ materially from those described in the forward-looking
statements. The forward-looking statements contained in this conference call
address the strategic business combination of Marvell and Galileo. The following
factors, among others, could cause actual results to differ materially from
those described in the forward-looking statements. The risk that Marvell and
Galileo businesses will not be integrated successfully; the costs related to the
business combination; failure of the Marvell or Galileo shareholders to approve
the business combination; inability to obtain or meet conditions imposed for
governmental approvals for the merger; inability to further identify, develop,
and achieve success for new products, services, and technology; increased
competition and its effect on pricing, spending, third-party relationships and
revenues; as well as the inability to establish and maintain relationships with
commerce, advertising, marketing and technology providers.

     For other factors that could cause the company's results to vary from
expectation, please see the risk factors section of Marvell's registration
statement on Form S-1 relating to the company's initial public offering, and the
company's quarterly report on Form 10-Q for the quarter ended July 31st, 2000,
and see Galileo's statement, including its 20-F statement on file with the SEC.

     Investors and security holders are advised to read the proxy statement and
prospectus regarding the business combination, when it becomes available,
because it will contain important information. The proxy statement and
prospectus will be filed with the SEC by Marvell and Galileo, and may be
obtained from either company or through the SEC.

     Before I turn the call over to Sehat, I'd like to briefly discuss the
terms, accounting treatment, and financial impact of this merger. Under the
terms of the merger agreement, Marvell will issue .674 shares of its common
stock for each common share of Galileo stock, in a tax-free stock-for-stock
exchange. As a result, Galileo's shareholders and optionholders will receive
approximately 32.9 million shares of Marvell common stock, or 25% of the diluted
ownership of the combined companies, in exchange for all the shares and options
of Galileo. Based upon Marvell's closing price yesterday at $81.75 per share,
this represents a valuation of approximately $2.7 billion.

     Subject to the approvals of both Marvell's and Galileo's shareholders, and
to the U.S. and Israeli regulatory approvals, we expect to close this
transaction in the first calendar quarter of next year.

     We will account for this merger as a purchase. On a pro forma basis,
excluding the amortization of goodwill and stock compensation expense, we expect
this merger to be accretive to Marvell's EPS upon closing.

<PAGE>   3

MARVELL TECHNOLOGY GROUP -- October 17, 2000                              Page 3


     We do not expect any material changes to the tax rates of either Marvell or
Galileo operations.

     Now, let me turn the call over to Sehat.

SEHAT SUTARDJA

     Thank-you, George.

     Thank-you all very much for being here with us today. This transaction is a
win-win for customers and shareholders alike. It combines two of the
best-in-breed broadband technologies, a leader in physical layer communications
silicon, with a leader in higher-layer packet processing, switching, routing and
system management.

     Our products, therefore, are highly complementary. For those of you who are
new to us, let me offer the following very simple explanation of our respective
technologies.

     Every communications systems has two basic parts. At one end, you will find
the physical layer, which provides the interface to the real world, analog
signals. In the end of the systems, integrated circuits... in the end of the
systems, integrated circuits convert the analog signals into digital
information, and Marvell is the leader in proving these high-performance silicon
solutions.

     The signal then moves from the physical layer to the other end of the
system, where the digital information is modified, processed, and ultimately
routed to its next destination. Galileo is the leader in providing the silicon
solution for this end.

     By bringing our companies together, we will be able to deliver end-to-end
solutions. In fact, we have partnered closely with Galileo in the past, and know
first-hand how compelling the combination is for our customers.

     Our combination with Galileo will accomplish four objectives. First, as I
mentioned, it allows Marvell to offer complete end-to-end silicon solutions to
communications equipment vendors.

     Second, the combined company will be the only supplier shipping both
gigabit Ethernet physical layer devices, and switching ICs in volume production
today. The merger immediately expands our market opportunity in the
rapidly-expanding gigabit Ethernet market.

     Third, it accelerates our entry into new markets. Together, Marvell and
Galileo posses the core semiconductor technologies required for future
communications systems in the MAN, WAN, and SAN.

     Fourth, it creates a communications IC powerhouse with the critical mass to
bring high-speed broadband quality solutions to the communications market.


<PAGE>   4

MARVELL TECHNOLOGY GROUP -- October 17, 2000                              Page 4


     Let me spend a few minutes describing the opportunity that Avigdor and I
see for our products. Marvell is an industry leader in high-speed analog mixed-
signal and digital-signal processing, using standard CMOS processes. We have
proven the strength of our core technology by being the first to market with a
production physical layer device, the Alaska _____ family, for gigabit Ethernet
over copper that achieves power levels that will make broad deployment of
gigabit Ethernet over copper a reality.

     We believe there's no vendor with better mixed signal and digital signal
processing technology for physical layer applications. We will continue to
leverage our core technology to enter new markets in the future.

     Similarly, Galileo is an industry leader in providing communication systems
on silicon. Its core expertise complements Marvell. It provides... complements
their handled packet processing, switching, and system management. Galileo
provides highly-integrated system controllers that are critical to customers
such as Cisco, Ericsson and Lucent. It introduced the first merchant single-chip
Ethernet switch in 1996. It continues to innovate. And the new product families,
the (GalNet 2-plus), the GalNet 3, and the Horizon product families, are gaining
momentum, with over 100 design wins to date. The Galileo components complement
our physical layer technology by providing layer 2, 3, 4 and 5 and packet
processing functionality.

     The immediate opportunity for our combined product portfolios is in the
rapidly-expanding gigabit Ethernet market. Marvell is uniquely positioned to
meet that need.

     To give you an idea how big this market is, __________ just released
another upward revision for this market. The number for gigabit Ethernet ports,
and 10 gigabit Ethernet ports shipped world-wide is expected to reach
approximately 75 million in 2004, a cumulative annual growth rate of over 100%
from year 2000.

     Starting next year, the number of copper gigabit Ethernet ports shipped is
expected to exceed the number of fibre ports. We intend to capture a substantial
share of that opportunity. Today, our combined company will be the only supplier
shipping both gigabit _____ and switching ICs in volume production. And as a
result, we will be uniquely positioned to provide end-to-end solutions to the
market. Marvell _____ physical layer device enables extremely high density in
the chassis. The GalNet 3, for example... the GalNet 3 switching family, for
example, provides advanced features such as guaranteed availability of service,
first floor accounting, and supports virtual private networking.

     These products will address the requirements of the most sophisticated
enterprise data center and network-edge layer 3 routers, and serve as a powerful
platform for the emerging MAN service providers, ASP, ISP and storage service
providers.

<PAGE>   5

MARVELL TECHNOLOGY GROUP -- October 17, 2000                              Page 5


     Our two companies also share a common vision for next-generation products,
particularly in addressing the 10 gigabit Ethernet and network storage utilizing
Ethernet and IP technologies, where the opportunities are very large.

     Now, let me turn it over to Avigdor, who I am pleased to announce will be
joining the Board of Directors of the combined company, along with Manuel Alba.

AVIGDOR WILLENZ

     Thank-you, Sehat.

     We are thrilled to join with the Marvell team to create a true
communications semiconductor leader. I think those of you who know us well
understand that we have always valued our independence. We are an innovator in
this business. We offer creative architectures to solve our customers' system
needs. Our new product families demonstrate this approach. The reception in the
market for our new controller and GalNet 3 family has been phenomenal. We're at
an inflection point. The benefit of forward thinking is evident through our
design wins with tier-1 communications equipment providers.

     Clearly, we could have remained independent. We chose to merge with Marvell
because it's the leader in one of the most critical areas of communications
systems, high-speed signal processing. At Galileo, we are in a unique position
to evaluate many different company technologies. We have built evaluation boards
and reference designs with many physical layer silicon vendors. Marvell's team
consists of world-class design engineers solving the very difficult challenges
of extremely high bandwidth communication. We've worked with virtually everyone
in this area, and we think that Marvell has the best team and the best solution.

     Let me spend a few minutes describing why these core technologies -
Marvell's physical layer combined with Galileo's packet processing - are so
essential to the future.

     The communication infrastructure is changing to accommodate the explosion
of traffic carrying voice, video and data. Carriers and service providers are
deploying new optical technologies to dramatically increase the bandwidth and
flexibility of the network. These new networks are capable of delivering
terabits, and in the not too distant future, even petabit data transfer rates.

     Terabit performance levels require a different approach to system design.
Single semiconductor devices can no longer deliver all the functionality at 10
and 40 gigabits per second speeds. To match these higher speeds, semiconductor
designers must deliver chip set architectures, not stand-alone chips, with
multiple circuits each operating at multi-gigabit clock rates that can work
together to handle tasks that today might be performed by a single chip when the
requirements are lower speeds.

<PAGE>   6

MARVELL TECHNOLOGY GROUP -- October 17, 2000                              Page 6


     To us the challenge and the opportunity is clear, delivering complete
systems and silicon capable of operating at the wire speeds of next generation
high-speed networks. This combination opens this door. Together our companies
know how to make chips that run at these speeds. Marvell is the strongest
merchant silicon supplier at solving these challenges in the widely-available
signals process, which is essential to integration. And also of importance, this
combination creates a company that knows how to put its high-speed chips
together in subsystems that work with one another.

     The immediate synergies in gigabit Ethernet are clear, but as I hope you
can see, the combination is even more compelling for the long term. Our
companies share a common vision for next generation product, particularly in
addressing 10- and 40-gigabit technologies, and network storage utilizing
Ethernet and IP technologies. The possibility set created by combining Marvell
and Galileo is immense.

     As Sehat indicated, I will be joining Marvell's board, together with Manuel
Alba, the president of Galileo, and look forward to continuing with the Marvell
team for a long time.

     Now, let me turn it back to Sehat.

SS:  Thank-you, Avigdor. This combination gives us the critical mass to pursue
     these opportunities. Through the merger we will double our existing sales
     force, double the number of design engineers, and roughly triple the number
     of our application engineers. Clearly, we will have the muscle to market
     our powerful portfolio of products aggressively, and develop new ones.

     By way of organization, Avigdor will continue to run Galileo as a
subsidiary of Marvell. As mentioned earlier, we are delighted that he and Manuel
will be joining Marvell's Board of Directors. The Galileo reporting structure
will remain essentially unchanged. Galileo will continue operations in Israel
and San Jose.

     I would now like to open the call to questions about the merger, then we
will devote the rest of the call to Galileo's third quarter results. Operator,
please take the first question.

QUESTIONS AND ANSWERS, PART I

OP:  Thank-you, sir. Ladies and gentlemen, we will now begin the question and
     answer session. If you have a question, please press the "1" followed by
     the "4" on your telephone. You will hear a three-tone prompt acknowledging
     your request. If your question has been answered and you wish to withdraw
     your polling request, you may do so by pressing the "1" followed by the
     "3". If you are on a speaker-phone, please pick up your handset before
     entering your request. One moment, please, for your first question.

<PAGE>   7

MARVELL TECHNOLOGY GROUP -- October 17, 2000                              Page 7


     Your first question is from Dan Niles of Lehman Brothers. Please proceed
with your question.

DN:  Thank-you. Maybe the first one's for George. You talk about, I think, a
     little bit, in the press release, that you expect this to be accretive upon
     closing. I was wondering if you could go through that a little bit. Is that
     just driven just by... I mean, how much cost savings are you assuming in
     that? Can you give a size of magnitude to that, George, if you would? And
     for, maybe, you, Sehat, could you kind of give an idea of... I mean, are
     the two companies, to some extent, even prior to the merger, going to be
     working together on joint products, and when could you see, sort of, the
     first joint product coming out of the combination, I guess?

GH:  Let me address the first part of that, Dan, and Sehat will respond to the
     second half. Actually, in our statement relative the accretive effect of
     the transaction, we basically just took a look at the public models that
     are out there for us that are being carried by the various analysts that
     cover both ourselves and Galileo, and basically just combine those two. And
     the net effect, from an accretion standpoint, is for calendar year 2001 we
     expect the accretive effect to be approximately 51%.

          Now, there could obviously be - and we do expect - significant
     additional synergies that will take place, that should help both... should
     help our combined financial result. We haven't quantified that at this
     point, yet, but we do expect it to be substantial.

DN:  Okay. And I guess I'm not going to get you to define "substantial" any more
     on this call, huh, George?

GH:  Well, we think that there is... we are kicking around some numbers that we
     think there is at least, from a top-line standpoint, a minimum of $25
     million more of combined revenue that we could easily go after, but we
     haven't had enough time yet, Dan, to really go into a lot more detailed
     analysis of that.

DN:  Okay. Well, that at least gives me a good starting point.

SS:  Regarding the R&D, the future joint projects between Marvell and Galileo, I
     want to address that by addressing the synergy of combining the two R&D
     teams together. This, actually, instantly creates a very powerful company
     that has tremendous knowledge and expertise in all areas of communications
     systems, from the physical layer all the way to the system software, packet
     processing, routing, switching, and system management functions.

          So, obviously, one of the things that we'll be working closely
     together is in the area of 10-gigabits, 40-gigabits, delivering system
     solutions that can deliver terabits per second in the throughput. But the
     immediate benefit there will be... that we're real excited


<PAGE>   8

MARVELL TECHNOLOGY GROUP -- October 17, 2000                              Page 8


     about, is the fact that the two combined companies will be the only company
     today, in the industry, that can ship volume productions of gigabit per
     second on a per port basis. So, we're talking about a system that will
     deliver multiple ports of gigabit switching. For example, a 12-port systems
     on a blade, even a 32-port system on blade types of solutions.

          Again, Galileo has shipped products, whether they're GalNet 2-plus, or
     GalNet 3, they are capable of supporting many, many ports of gigabit, as
     well as the older generations of 100-megabit solutions, on the same
     systems. We have shipped the companion gigabit five as well as the 100
     megabit five, all based on the most advanced mixed signal and digital
     signal processing technology in the business, as well as being implemented
     in the most advanced 0.18 micron seamless processes, delivering the
     lowest-power solution.

          So, in time, integrations will also make sense as we move to a smaller
     and smaller geometry. This is what, you know, obviously, we'll do in the
     future.

DN:  Great. Thanks a lot. Sounds like a great merger.

OP:  Your next question is from Brian Wu of Bear Stearns. Please proceed with
     your question.

BW:  Hi, guys. Quick question on, maybe, some of the other synergies outside of
     being able to provide an end-to-end solution. Are there any new customers
     that are now part of the combined entity that weren't there before? Were
     there any other synergies that you'd like to talk about?

SS:  Sure. If you have followed Galileo, you are aware that Galileo has about,
     what, 600 customers. We have about 60 customers. The combined companies
     will have a bigger customer base, as well as be able to deliver practically
     all the components, the critical semiconductors that are needed in the
     systems. So, we believe by combining the companies we will have all the
     solutions our customers need, as well as we will have a bigger team to
     solve some of the future needs of our customers, and be able to deliver
     future solutions sooner.

          To give an example, when we're talking about terabit per second
     systems solutions, as Avigdor mentioned earlier, the way we architect the
     system has to be different than the way people architected the system in
     the last couple years. While, in the past, people are dealing with server
     ports of, you know, 100 megabits per second solutions, maybe a few ports of
     gigabit, a single-chip solution may make sense. In the future in order, to
     build systems that can approach terabit per second, distributed processing
     is a must. And one way to approach... one way to solve this problem... a
     technology that's important to solve this problem is high-speed
     chip-to-chip interface

<PAGE>   9

MARVELL TECHNOLOGY GROUP -- October 17, 2000                              Page 9


     ___________ in a tens of gigabits per second surges technology, for the
     chip-to-chip connect as well as for backplane applications.

          So, by combining the two resources together, we will be able to
     address... deliver future products sooner.

BW:  And quick follow-up on... you said greater than 100 design wins. Was that
     for the combined three products, GalNet 2-plus, 3 and Horizon?

SS:  Yes.

BW:  Could you maybe break that out a little bit more? Like...

GH:  I think we don't have that. Galileo will address that a little more when
     they talk about their third quarter.

BW:  Okay. All right, thanks, guys.

OP:  Your next question is from Alan Septimus of Oscar Bruce. Please proceed
     with your question.

AS:  Thank-you. Most of my questions have been answered. Is there a walk-away
     price, termination price, whereby if Marvell's stock closes, during some
     pricing period, below it, that the transaction could be terminated?

GH:  No. At this point there is nothing contemplated. There's no collar involved
     in this transaction. There is now a fixed exchange ratio at .674, so
     obviously the value of the transaction now is based and keyed off of
     Marvell stock price, but no, there's no walk-away.

SS:  I want to add on this, as well. Actually, we've been working with the
     Galileo team for the last six months, doing the joint sales and marketing
     as well as future product development alignment. Along the way, when we
     work together, we all feel that we have the same synergies, we feel we have
     the same cultures, ambitions of the future, and both of us feel that this
     merger, that comes, actually, quite late... later on that we realized that
     what comes naturally as a win-win transaction for our customers and
     shareholders alike. So, we believe that this will be an extremely
     important, as well as successful transaction.

AS:  Okay. So that the price behaviour today, of course it could always be
     temporary. This would only influence... the only impact this would have is
     if shareholders were not to approve the transaction. But there's nothing
     inherent in the merger agreement that says the transaction doesn't move
     forward of Marvell's stock price falls.

GH:  Correct.

AS:  And in addition, are there any... is there just a general material adverse
     change clause in this contract?

GH:  Yes, there is.


<PAGE>   10

MARVELL TECHNOLOGY GROUP -- October 17, 2000                             Page 10


AS:  Okay. Thank-you very much.

GH:  Okay.

OP:  Your next question is from Alex Gauna of Bank of America Securities. Please
     proceed with your question.

AG:  Yes, I was wondering if I could encourage you to maybe talk a little bit
     more specifically about perhaps some Marvell platforms where Galileo might
     now have an opportunity to develop... design in its switching products, and
     vice versa, of those 500 customers perhaps some specific or star platforms
     or customers that might now be able to draw more heavily on Marvell's
     physical layer devices?

SS:  Are you asking for specific programs?

AG:  Yeah, I mean, your more exciting programs and design wins between the two
     companies, where you're going to be able to target the new entity's joint
     product line.

SS:  Okay, what actually we're seeing from the last several months of visiting
     the customers when we were doing the joint sales and marketing process, we
     saw a lot of opportunities for the future applications of highly scalable
     metropolitan area network types of applications for systems that can
     support hundreds of gigabits, supporting terabits per second types of
     solutions. So, obviously, these are the areas that we'll be working closely
     with the Galileo team to address the applications that are required by the
     customers. I mentioned earlier that there are immediate opportunities on
     the current gigabits types of applications, and we're seeing, for example,
     a lot of customers are demanding... asking for solutions that support, say,
     12-port or 24-port gigabit switching systems, and together we'll be the
     only ones that can provide this solution in the market.

AG:  Are there any... of these 12- and 24-port systems, are there any customers
     or potential customers you're willing to name at this time?

SS:  No, not at this time.

AG:  Okay. Thank-you.

OP:  Your next question is from Tom Vanbuskirk of Silverado Capital Management.
     Please proceed with your question.

TV:  Hi. I apologize for this. I had to leave the call for a minute. I hope you
     didn't already cover this, but could you just spend a moment and describe
     what the Israeli regulatory approval process is like?

GH:  Sure, we'd be happy to do it. The merger requires approval from various
     agencies within the Israeli government. The Investment Center is one area,
     and the Department of Corporations and so forth. But then Israel will need
     to review the contract, and then we will then submit it to the shareholders
     for their vote. We're not anticipating any problems

<PAGE>   11

MARVELL TECHNOLOGY GROUP -- October 17, 2000                             Page 11


     with it. It's a very straightforward merger, especially with Galileo being
     kept as a subsidiary of Marvell. It's very, very clean, so we're not
     anticipating - although you never know - we're not anticipating any
     regulatory issues.

TV:  Is there any sort of a time-line that goes with that?

GH:  As we indicated earlier, we expect to close the merger in the first quarter
     of next year. First calendar quarter of next year.

TV:  All right. Thank-you.

OP:  Your next question is from Nathaniel Cohn of Goldman Sachs. Please proceed
     with your question.

NC:  Yes, thank-you. I was wondering, I guess, following along a prior question
     regarding design wins, can you talk at all about the number of design wins
     that you might have together, and if it's possible, could you talk a little
     bit about your momentum that you're seeing, currently, on the gigabit
     products, on the physical layer? Thanks.

GH:  Okay. Let me tackle that a little bit. I think... let's first of all review
     where Marvell stands. Those of you that are familiar with us know that we
     updated you, on our second quarter earnings call, that we had achieved
     initial design wins in both the NIC part of the gigabit market as well as
     in the top tier-1 switching market. We had also been successful in gaining
     design wins there including Cisco.

          We've been making a lot of progress this quarter. You probably all are
     aware that Intel has announced their Pro-1000 network interface card. They
     were showing it at the recent Interop Trade Show. That product is based on
     their MAC and the Marvell 5, and we are shipping volume production to them
     here in our third quarter, which ends at the end of October.

          We've also begun some initial production shipments to some of those
     design wins, you know, with the major switching guys that I had just spoken
     to.

          We are making a lot of progress. We now have significantly increased
     the number of design wins that we previously reported at the end of the
     second quarter. We'll be reporting our earnings in approximately one month,
     and we will give you a more significant update relative to actual numbers
     of design wins at that point, but suffice it to say that there has been
     substantial progress made since our Q2 report.

          It's probably too - as Sehat mentioned a minute ago - it's probably
     too early, yet, to specifically indicate where the combined products have
     already been successful. We have had a co-marketing program going on now
     for several months, that we believe is yielding some success, and it's a
     little early yet to say specifically what customers are moving forward with
     that, but we now expect that rate to increase significantly as we now

<PAGE>   12

MARVELL TECHNOLOGY GROUP -- October 17, 2000                             Page 12


     offer the one complete end-to-end solution to the market. So, I think, as
     we move forward - and possibly by the time of our call, I'm not sure yet -
     we will be able to update you with more facts about that.

NC:  Okay, great. And then I guess one following question. Can you talk a little
     bit about the premium that you're paying, and if you could put that into
     perspective, that'd be great.

GH:  Sure. Well, as you know, the equity markets have been extremely volatile.
     Over the course of these negotiations, the premiums moved around
     significantly. So, if you focus exclusively on the premium involved the
     markets, you can potentially miss out on important strategic opportunities
     like this one.

          We believe that the financial and strategic merits - as Sehat
     described, and Avigdor's commented on - of this transaction justify Marvell
     having Galileo's shareholders own 25% of the combined company. We believe
     the financial metrics of the transaction are compelling. As I mentioned
     earlier, there is significant EPS accretion of greater than 50% in calendar
     2001. There's margin enhancements from the combination of the two
     companies, and an improved growth profile over the long term.

          And another key point, we believe, is that Galileo's business is at an
     inflection point. The strategic benefit of the transaction will only serve
     to accelerate the momentum. So, we're very comfortable with the financial
     metrics of this transaction.

NC:  Okay, great. Thanks, and congratulations to both companies.

GH:  Thank-you.

AW:  Thank-you.

OP:  Your next question is from Charlie Glavin of Credit Suisse First Boston.
     Please proceed with your question.

RE:  Actually, it's Regina Eberhard calling in for Charlie. I was wondering...
     I've got two questions. One is related to Galileo's system controller
     business, and the road map in that business. Given the merger, can we
     assume that your product road map, Avigdor, is going to continue in that
     area?

AW:  Yes, absolutely. The system controller business is extremely strong for us,
     it's a high margin business, very defensible. Basically we're almost alone
     in that market at this point, and we're definitely planning to continue and
     enhance, and I think that this quarter we're going to announce a
     significant new product in that area.

RE:  And then my second question is more a financial one. You guys... Galileo
     obviously has a very nice tax benefit from Israel, and it sounds like you
     guys expect that to continue. Is that correct?

<PAGE>   13

MARVELL TECHNOLOGY GROUP -- October 17, 2000                             Page 13


GH:  That is correct, yes. And the reason that's... the way we're again,
     accomplishing that is that Galileo is going to be operated legally and
     operationally as a wholly-owned subsidiary of Marvell, and that will permit
     us to protect and keep intact the current tax structure for Galileo, and
     combine it with the very... also efficient tax structure or Marvell.

RE:  So, what are... is it too early to say kind of what Marvell's corporate tax
     structure will look like, kind of looking ahead into 2001?

GH:  Sure. Well, we have... right now we are recording 25% tax rate on our
     earnings, and we are... we actually believe, in the future that rate could
     be substantially lower than 25%, but we believe at this point it's prudent
     to continue providing a provision at that rate until we're sure we have
     cleared any regulatory challenges to our structure, but we feel very, very
     comfortable that we, in the long term, probably will be less than 25%.

RE:  Thanks a lot, guys. Good transaction.

GH:  Thank-you.

OP:  Your next question is from Terry Ragsdale of JP Morgan. Please proceed with
     your question.

TR:  Yeah. I probably have to admit I don't know probably as much about the
     Galileo product line-up as I maybe should. Could you walk us through what
     the gigabit switch products are at Galileo, now, which ones are in volume,
     etc., and what kind of switch port densities you can reach with them?

AW:  You're asking specifically only about the switches?

TR:  Yes.

AW:  Okay, so, on the switches our position is not at the low end of the market,
     but at the medium- to high-end, and our switches are sold into applications
     that require modularity and high port density, and applications that are
     based on chassis-based systems for reliability and high performance. What
     you can get with our product, it's actually with our chip sets, it's the
     only merchant silicon in the market that you can build of 256 ports of
     10/100 and 32 ports of gigabit shipping today, and that is in layer 2 and
     layer 3 and layer 4.

          So, we have several families that are based on the same platform of
     switching, which is the GalNet 2, GalNet 2-plus and GalNet 3, where the
     differences are in the level of integration of memory between the GalNet 2
     and GalNet 2-plus, and some layer 3 functionality, but the major difference
     is to the GalNet 3 that brings features that are applicable to service
     providers and metropolitan area networks, like building a service

<PAGE>   14

MARVELL TECHNOLOGY GROUP -- October 17, 2000                             Page 14


     level agreement with customers, rate-shaping, and so on. Those are very
     high-end products in the Ethernet market.

          On top of that, the GalNet 3, we have some new products in the GalNet
     3 family that are just starting to ship, that support packets over SONET,
     and ATM OC/12 rates, combined with Ethernet. So, you can build and mix
     and match with our architecture anywhere from 8-port to 256-port, gigabit
     and 10/100, layer 2, layer 3 and layer 4, packet over SONET, ATM and
     Ethernet. So, actually, it's the only family of merchant silicon that has
     this capability that it's shipping in volume production, with some of the
     components over a year and a half.

TR:  And can you walk me through, a little bit, what the landscape is, in terms
     of merchant silicon for gigabit switching. I mean, you mentioned you're the
     only one out there. To what extent are OEMs interested, still, in
     developing their own switch fabrics for gigabit, and to what extent are
     they going merchant?

AW:  Right. Still some significant amount of the switching in gigabit is still
     done in-house. When you look at the leading companies at that space,
     they're still designing their own silicon. But our opinion, and especially
     now with the combination with Marvell, and the integration that we are
     going to bring between the switch packet processing and the physical layer,
     and the service capability, the chip-to-chip connectivity and backplane
     technology that Marvell brings, we will be able to build... and also, by
     the way, the speeds, the library that Marvell brings, and their capability
     of running at gigahertz speeds, those are very, very significant to the way
     we're going to build our next generation GalNet family, that you saw some
     press releases - at least, there was one press release - starting to talk
     about the GalNet 4 family.

          That technology, and those capabilities, we believe to process at
     terabit speeds, you will not be able to do it in-house. Only maybe some of
     the very leading companies like Cisco and so on, that have very, very big
     design teams, will be able to continue and develop at .18, .15 and .10
     technologies, but we think that because of the mixed signal and DSP
     capabilities, they will find our solution extremely compelling.

          People just don't understand what's going on, because I'll give you
     one example. If in .35, to do a product, the NRE - the non-recurring
     engineering - was $85,000 for the mask set. At .15, it's half a million
     dollars, and at .10, it will be probably $1.5 million. So, if you're just
     doing it for in-house, you have to amortize it on the number of products
     that you sell, versus buying it from a chip vendor. But that's not the
     biggest problem. The biggest problem is the knowledge and verification of
     system-level silicon. Galileo today is by far one of the five leading
     companies in formal verification.


<PAGE>   15

MARVELL TECHNOLOGY GROUP -- October 17, 2000                             Page 15


          Formal verification enables us to do... I mean, you have it in Bell
     Labs, I think Intel has it, IBM... IBM is a key developer of this
     technology, and we were their first, and I believe the only alpha site
     world-wide for this technology. We even co-wrote some patents with them on
     that. This is extremely significant on the way that you'll be able to do
     system-level silicon in the future.

          So, just to summarize a long answer, we believe that, going forward,
     there will be just no way that many of the system companies will be able to
     continue and do effectively system-level silicon.

TR:  And when do you see that sort of transition happening, where the in-house
     efforts run out of gas and they've got to go merchant. Is that a year out,
     is it three years out?

AW:  It's not a year out, and not... it's a process that is happening. If you...
     one day we'll meet, and I'll show you our business plan we wrote in '92. We
     wrote exactly that, and what we see over the years, since the time that we
     started in '93/'94, and especially '95 when we entered into the
     communications space, moving from the printed space, we saw this
     transitioning happening slowly, slowly, and companies that said, in the
     past, `No way that we'll outsource our core competency of building our own
     silicon, our own ________ to our largest customers,' and we just see this
     as a process that is happening more and more into the future. It's not that
     it will happen next year, but it's happening constantly. Every month, every
     quarter you see companies that fail or were unable, because of focus, and
     enough energy to put behind those problems, were unable to provide the
     solution and are going out to merchant silicon, and that's why you see
     some of the success of the merchant silicon in the communications space.

TR:  I have two other quick ones, and I promise I'll go away after that. The
     first one on integration. When you mentioned integration, it sounds like
     what you're talking about there is kind of - and again, using some of your
     words - changing the system architecture a little bit. It's not necessarily
     putting the layer 1 on the same piece of silicon with the layer 2 and
     above.

AW:  Right. You're absolutely correct. One thing is really putting the layer 1
     and layer 2, but it will not... it's not the right partitioning in every
     case. In many cases, and for us, the people that try to drive the
     technology, and when you really look at the space, the optics are creating
     a big problem for the electronics, for the semiconductor, for the system
     level that is not run by optics. So, you have the...

[End of Side A]

<PAGE>   16

MARVELL TECHNOLOGY GROUP -- October 17, 2000                             Page 16


[Start of Side B]

AW:  ...and what the technology that Marvell brings us that is, for me, one of
     the most exciting parts, even more exciting than the integration of the
     level 1, is the ability to do this incredible chip-to-chip connectivity and
     backplane infrastructure to build a distributed architecture to address the
     needs of the optics at terabits and beyond. And we are, from our first
     silicon in '96, believed in distributed architectures. So we mastered this
     space. And we have competitors that try to enter this space, to remind you,
     like, TI and PMC Sierra, with also similar architectures, with distributed
     processing, that couldn't compete in the end and pulled out of the market.
     We, over the years, not recognized by Wall Street, were able to deliver to
     our customers this type of very, very sophisticated architecture. What
     Marvell brings us, now, is really the capability to go to the next step -
     and a big step - into the 10 gigabit speeds and above.

TR:  All right. And the last one is, two years out, who do you see being your
     most serious competition in this area?

AW:  I think that... I found out over the years that always the most serious
     competition is ourselves. It's the ability, really, to execute and deliver
     on time clean silicon. That's the number one. But still I think it's still,
     to some respect, in-house, that in my opinion, over the long term, will
     have to go more and more to merchant silicon. Maybe not on everything. I
     think there will be very interesting and appealing combinations of FPGAs. I
     know that the market now downgraded FPGAs, but I think, long-term, the
     solution will be a solution that really fits the way that our vision is on
     partitioning and system architectures, which means many of the
     functionality that is already standard will run and merchant silicon at
     very high speeds with high write integration with boards that will
     enable you to interface to FPGA, do very high-speed FPGAs for
     differentiation by the customer.

          So... and, of course, to a processing like standard processors, and
     network processors, that are interfacing to standard merchant silicon for
     the fabric, and the standard packet processing.

          So, in-house will be, to some degree, but you probably can name some
     of the significant communication merchant silicon suppliers that are
     potentially addressing - that are addressing - similar markets, but they
     could potentially be a competitor. But, unlike the PC market, this market
     is so diverse, you have so many issues, so many problems, and so many
     opportunities, now, in this convergence of the complete change in how
     telephony is done, the services that you'll be able to get with soft
     switches, from how storage is done on the network. There are so many
     segments and sub-segments that

<PAGE>   17

MARVELL TECHNOLOGY GROUP -- October 17, 2000                             Page 17


     need, desperately, solutions today, that in many cases it will not be a
     direct competition. It's not the DIRA market, or the Pentium market, or
     other markets that you basically can replace the function by another. It's
     multi-dimensional competition on features.

          So, we can probably meet certain segments in the market better, and
     some of our competitors will meet other segments.

TR:  Great. Thank-you.

GH:  Thanks, Terry.

OP:  Your next question is from Steven Angeli of Wellington Management. Please
     proceed with your question.

SA:  Hi. We talked about, earlier, about a limit in terms of price, but the
     stock in the past, Galileo, specifically, speaking as a Galileo
     shareholder, has traded and is being sold, it seems, at a significant
     discount to comparable companies in the industry. And as you said yourself,
     Galileo's at an inflection point in its fundamentals. Obviously I realize
     you couldn't predict what was going to happen in the stock price, but the
     purchase premium has gone from 75% to, you know, a small 13%. Again,
     realizing you couldn't have predicted exactly what was going to happen, but
     is this a deal that you will still do no matter what the price ultimately
     turns out to be?

AW:  We're really looking at the long-term. I mean, market fluctuations, we
     really look at our stock over the last three years, I think we brought
     tremendous value to our shareholders by not looking just at the stock price
     but looking at strategies, by looking at the way we believe the business
     needs to be done.

          And our story was always complex. That's one of the reasons that we
     never got the multiples that our comparables enjoyed. I think this story is
     complex, but we'll make a major effort to explain it, and I think that by
     explaining it and showing the results, long-term there is no question in
     our minds, the two managements of this company, that this will be a
     tremendous success.

SA:  Thanks.

SS:  Yeah, I agree with Avigdor.

OP:  Your next question is from Michael Emerald of Longfellow Investments.
     Please proceed with your question.

ME:  What U.S. approvals will you need, and which approvals form the critical
     path to get you out to the first quarter?

GH:  We will, you know, go through the standard U.S. regulatory approvals, as
     well as submitting our proxy and S-4 to the SEC for their review and
     comments, and then

<PAGE>   18

MARVELL TECHNOLOGY GROUP -- October 17, 2000                             Page 18


     obviously distributing the proxy to shareholders, and then obtaining their
     vote. We believe we're in a reasonable position to accomplish that task
     within a 90-day period.

ME:  Thank-you.

OP:  Your next question is from Charlie Glavin, a follow-up. Please proceed with
     your question. [pause] Mr. Glavin, do you have a question, sir?

GH:  Operator, I think what we'd like to do now is move to Galileo, reviewing
     their third quarter results, and then taking their Q&A, and then we'll be
     more than happy, in the end, if there are still further questions
     regarding, you know, the combination, we'd be happy to address them, but we
     want to make sure we give Galileo appropriate time to review their results.

          So, with that, I'd like to turn the call over to Avigdor to review
     Galileo's third quarter results.

</PRE>