October 09, 2012 - 9:48am EST by
2012 2013
Price: 6.82 EPS $0.26 $0.00
Shares Out. (in M): 54 P/E 26.0x 0.0x
Market Cap (in $M): 368 P/FCF 0.0x 0.0x
Net Debt (in $M): 0 EBIT 17 0
TEV (in $M): 231 TEV/EBIT 13.0x 0.0x

Sign up for free guest access to view investment idea with a 45 days delay.

  • Semiconductor
  • Acquisition Target
  • Technology


The MIPS sale process is near a conclusion. There are currently only two viable low-power CPU architectures with significant adoption: ARM and MIPS. As such, MIPS is a rare asset that may be extremely valuable in the huge mobile device market. A DealReporter story dated Sept. 18 (“MIPS Technologies sale moves forward”) reported that 6 bidders are involved in the auction. Names mentioned include INTC and QCOM, as well as patent trolls such as ACTG and RPXC.

The intense competition in the mobile device market, with a number of parties with a variety of offensive and defensive reasons to own MIPS, is likely to lead to a very robust auction. MIPS’s current $235mm enterprise value is peanuts compared to the stakes in the mobile market. The MIPS bell cannot be un-rung and a transaction will happen in the near future. I believe there is very little downside at current prices with the upside option of a strong auction result.

The MIPS Architecture & The Broadcom Deal

The MIPS architecture was developed in the early 1980s at Stanford by John Hennessy and his team (Hennessy is currently the president of Stanford). MIPS CPUs were used in SGI workstations during the 1980s and 1990s, which lead to MIPS being acquired by SGI. MIPS was spun back out by SGI and began to license the MIPS architecture to third-party semiconductor companies. While the ARM architecture has clearly won in the mobile space, MIPS has traditionally been strong in consumer electronics and networking.

It is important to note that MIPS pre-dates the ARM architecture (MIPS’s earliest live patent pre-dates ARM’s by 6 years) and MIPS has ~580 patents surrounding CPU design. As evidence of the strength of MIPS’s IP, BRCM, a major licensee of both MIPS and ARM, paid $26.5mm cash plus other consideration for a license to MIPS’s patents in July after it was known that MIPS wanted to sell itself. The new license runs until the expiration of the last to expire of the patents being licensed to BRCM and does not include any pre-paid royalties (i.e., BRCM will continue to pay its current royalty rate on its MIPS-based products).

The general consensus among technology types is that the ARM and MIPS architectures are very similar and MIPS has some important patents around CPU design that other architectures likely infringe upon. BRCM’s behavior seems to support this view. If MIPS were to wind up in hostile hands, BRCM could over time transition over completely to ARM, but would still be vulnerable to litigation over its ARM products. Prior to this new license, BRCM’s MIPS license had another 6 years left, which is more than enough time to transition over to ARM, so BRCM was unlikely worried about patent exposure on its MIPS products.

In addition, the DealReporter article states that BRCM “sought a buyer for MIPS given fears that Intel could buy it and limit access to its technology”. Another unnamed MIPS customer “’aggressively’ sought buyers for MIPS to ensure it remains in ‘good hands’ so it will have access to its IP in the future”. I think MIPS was smart to do the BRCM deal since it gives other companies a comparable to look at to assess the value of MIPS’s patents. Also, it signals that if BRCM is worried about the patents as it relates to ARM, then everyone else should be worried too.

Who May Buy MIPS and Why

I think we can all agree that the mobile device market is huge and that the competition is intense. The table below summarizes a list of possible MIPS purchasers and the reason(s) for their interest in MIPS.

Intel (INTC)

I was pleasantly surprised to see INTC mentioned in the DealReporter article because if they are really involved, the likelihood of a strong auction result increases. I can think of three good reasons why INTC would want to own MIPS.

1) Despite INTC’s protestations, there is a good argument to be made that x86 is inherently less power efficient than ARM or MIPS. The x86 architecture uses variable length instructions that do not self-synchronize. This means that you can start reading at position 0 and get one valid set of instructions or start reading at position 1 and get another completely different set of instructions. As a result, the instruction decode circuitry in a x86 chip is complex and can almost never be turned off. For example, the decode front-end in an Atom chip can account for 20% of total power consumption. Owning the MIPS architecture would be a very cheap insurance against x86 never getting low power enough to effectively complete against ARM.

 2) Armed with the MIPS patents, INTC can use litigation to disrupt the ARM ecosystem both to attempt to gain share in the mobile market and to prevent ARM from moving into the server/notebook market. Although INTC has finally managed to get its Atom SOCs low-power enough to be used in a phone (i.e., Motorola’s RAZR i), INTC’s hardware partners or potential partners are unlikely to drive meaningful volume. Apple and Samsung (combined 40% of unit volume in 2011 and growing) are committed to their internal custom ARM silicon and thus closed to INTC. Outside of Apple and Samsung, the other material chunks of market share (HTC, NOK, Huawei, ZTE) are almost exclusively QCOM sockets. INTC would have to displace QCOM to gain any meaningful share, which will not be an easy task given the ecosystem around ARM. The threat of patent litigation may be an effective stick in getting those vendors to switch. This may explain QCOM’s appearance in the list of bidders in the DealReporter article.

 ARM and its partners are attempting to invade INTC’s traditional server and notebook markets. The release later this month of Windows RT will bring Windows to ARM chips (the Microsoft Surface tablet/notebook is ARM-based). To the extent that Windows RT cannibalizes x86 Windows notebook sales, INTC will be losing sockets to the ARM ecosystem. The ARM ecosystem is also seeking to move into the server market. Samsung is developing ARM-based server chips (and hiring server personnel from AMD). MRVL, AMCC, and Calxeda are also building ARM server CPUs. Patent litigation may be a weapon to help INTC defend its traditional PC and server markets.

 3) Prevent the MIPS patents from falling into the hands of a hostile party that may sue INTC.

 Patent Trolls

The patent trolls add an interesting dynamic because it appears no one wants the MIPS patents to fall into a troll’s hands. A troll’s interest obviously stems from the large ARM ecosystem that would be vulnerable to litigation. Again, BRCM’s actions in taking out the MIPS patent license lend credence to the view that MIPS patents can be an offensive weapon.

The two trolls mentioned by name, ACTG and RPXC, have the financial resources to acquire MIPS. ACTG has a $1.4b market cap and $430mm of cash. RPXC has a market cap of $610mm and $220mm of cash.

ARM Holdings (ARMH)

While ARM does not need the MIPS architecture, it obviously has a strong interest in protecting itself and its ecosystem partners from patent litigation. The MIPS IP portfolio owned by INTC or a troll could cause many headaches for the ARM ecosystem.

The DealReporter article stated a MIPS customer “found two companies willing to move forward with a bid, adding that he believed the interested suitors are ‘not from Asia or the Americas’.” Europe would count as not Asia or the Americas and the two candidates that pop into my head are ARM and Imagination Technology. I imagine the conversation with ARM went like this:

ARM Licensee: “Hello, uhm… do you know MIPS is selling itself?”
ARM: “I do apologize, what did you say? I was eating this delicious crumpet and drinking my tea. Did you know we just hosted a brilliant Olympic Games?”
Licensee: “MIPS is selling itself. You need to go buy MIPS so INTC or some troll don’t get it and sue us all.”
ARM: “Bollocks. Right, let us look into it.”

ARM has a $12.8b market cap and $600mm of cash.

Imagination Technologies (IMG.L)

Imagination is an interesting candidate. The company is a licensor of GPU IP (its PowerVR technology). Both INTC and AAPL took large stakes in Imagination to prevent its IP from falling into the wrong hands. ARM has entered the GPU space with its Mali GPU and has found success (for example, the Samsung Galaxy S III uses a Mali GPU). Imagination could use MIPS’s assets to offer a competing CPU architecture along with PowerVR. For example, Imagination has a stated intention to expand into the home consumer market, where MIPS is traditionally strong.

Imagination has a $2.2b market cap and $100mm of cash.

Large MIPS Customers (BRCM, CAVM, CSCO, MCHP, etc…)

MIPS’s customers certainly have a strong interest in MIPS’s IP remaining in friendly hands, but I find it unlikely any one of them would acquire MIPS outright, even with the mention of QCOM in the DealReporter article. BRCM has already taken out its insurance policy. The only structure that would make sense to me is some sort of syndicate of customers that buys the IP and maintains the status quo. Another reason a customer may buy MIPS, and perhaps this is what QCOM is thinking, is that a large semiconductor company does not want to be completely beholden to ARM.

Loongson (China)

The Chinese are developing their own CPU architecture to lessen their reliance on the West. You can read about it here: http://www.extremetech.com/computing/127791-china-plans-national-unified-cpu-architecture

Loongson is essentially China’s national CPU company and its architecture is based on MIPS. Loongson is a MIPS licensee. I think China would like to (and indeed has tried to, if reports are to be believed) buy MIPS, but obviously if the Chinese were to attempt to purchase MIPS, the U.S. government would give them the straight arm seeing that MIPS microcontrollers help run things like F-16s.

Risk & Downside

The major risk here is that no deal materializes, but I would put a low probability on that. This truly is a bell that can’t be un-rung. MIPS’s new license revenue has collapsed to nearly zero ($1.5mm in the most recent quarter ex-the BRCM deal) due to customer uncertainty. Customers are unlikely to commit to the MIPS architecture absent a resolution where MIPS is no longer independent. Why would you trust your product roadmap to a company that has already shown a willingness to go far down the strategic alternatives path? Given that MIPS is no longer viable in its current form, I think it’s unlikely that nothing happens, so the risk is in the structure of the transaction and the fact that an acquirer knows that MIPS has to eventually hit a bid.

I think the worst transaction structure would be for MIPS to be broken up and sold piecemeal. MIPS’s two main assets are the patent portfolio and the design IP business (which generates MIPS’s $50mm annual royalties).  Maximizing value in a piecemeal sale of these assets would require a specific combination of buyers. For example, INTC buying the patents and Imagination buying the design IP business since both are natural allies against ARM and can fully utilize the assets.

Looking at the piecemeal sale/liquidation scenario as the downside, the current market price implies a value of $77mm for the patent portfolio.

Cash: $137mm
Other Assets (AR, PP&E, etc…): $8mm
IP Design (PV of Royalties): $180mm (currently $50mm of annual royalties)
minus Wind-down Liabilities: $35mm

Total: $290mm
Current Market Cap: $367mm

Implied Value of Patents: $77mm (2.9x BRCM deal)

Let’s look at the $77mm implied value in the context of the $26.5mm BRCM deal. BRCM is far from ARM’s largest customer and this sophisticated party was willing to pay for peace in the patent litigation valley once it heard MIPS was for sale. It’s fairly easy to name 4 large ARM licensees that should be willing to pay what BRCM did: QCOM, TSMC, Samsung, and TXN. Can we name 4 more large ARM customers? Sure -- NEC, STM, ZTE, and AAPL. Obviously, this list can get quite long, but the point is the BRCM license should be a key comp for both a troll and the ARM licensees with potential litigation exposure. The point is even in the breakup/liquidation scenario, the downside at current prices is limited if the BRCM comp is to be believed.

The other main risk is that any potential acquirer knows that MIPS has to hit a bid since the company is no longer viable in its current form. I believe this risk is mitigated by the number of bidders involved and their diversity of reasons for buying MIPS. Absent collusion, which is highly unlikely given the adverse interests of the trolls versus the industry and the intense competition among industry players, it should be easy for another bidder to top any lowball offer.


Valuation is difficult since MIPS’s value on the numbers as a standalone is significantly different than the value of its assets in another company’s hands. If I am INTC, would I be willing to pay $10/share ($427mm net of cash) to hedge against the intrinsic power disadvantage faced by x86, litigate against the ARM ecosystem, and protect myself against IP litigation? If I am ARM, would I be willing to pay $11/share to protect my ecosystem from a credible threat of patent litigation? I think it is an easy yes for both companies.

So in conclusion, we have all the ingredients for a very robust auction.

1) A rare asset: MIPS and ARM are the only two viable low-power CPU architectures with significant adoption.
2) Huge markets at stake: Mobile, tablets, and potentially servers/notebooks.
3) Lots of interested parties: DealReporter article reports 6 interested bidders. Given how long this process has been going on for, I expect these 6 parties are serious and submitting bids.
4) Bidders have many offensive and defensive reasons to own MIPS.
5) Bidders have deep pockets.

I do not hold a position of employment, directorship, or consultancy with the issuer.
I and/or others I advise hold a material investment in the issuer's securities.


 Announcement of a strategic alternative(s).
    show   sort by    
      Back to top