We would love to hear your thoughts about our site and services, please take our survey here.
Rz105, many thanks for your insights. They are appreciated.
I've known for some time now the identity of the hyperscaler customer for which AWE are developing the connectivity products as I've heard them mentioned in several podcasts.
I 'm crazy enough to have attempted a revenue model projection and for 2024, it assumes about $5 million for the connectivity products. That's probably ultra conservative. Certainly the 1st generation products were qualified last year and are now in production with first deliveries later this year. Listening to the Q&As of recent webcasts hinted that some of the 2023 R&D spend is for the second generation products.
I think you're on the money with Tesla being the automotive customer. I really can't see who else it could be. I know Elon Musk was emphasising only a couple of days ago that Tesla are not just an automotive EV company and are much more about 'autonomy'. So whether, Elon's project is destined for success or failure, I couldn't say, but one thing is for sure, a lot of money is going to be invested into this.
Alphawave are throwing a lot of money into R&D, not only the $78.2 million they have expensed but also the $53.3 million they have opted to capitalised, meaning effectively 41% of this year's annual revenue value has been spent on R&D. I am a little bit nervous about the level of R&D spend here. This they have to do, to get the products ready for market, but leaves me questioning how much more is required before they can scale back on the R&D investment. If it continues at the present rate, a cash placing round might have to be considered, though I hope not. Alphawave's competitors incidentally are also committing similar amounts to development percentage wise. Marvell spent 34.4% of their revenue on it and Credo 41.7%, Rambus 34%. Both Marvell and Credo reported losses. Rambus would have done were it not for the fact they sold off the PHY IP to Cadence. Credo did a cash placing last year and it had zero impact on the stock price!
I'm not 100% sure that Alphawave are the only one's with connectivity IP on TSMC's 3nm. I know that Cadence has too.
I really don't think the revenue ramp up need be in doubt. In about 36 months from now they should be able to report revenue somewhere in the order of $700 million to $1 billion with revenue split roughly equally between the three main streams: IP licensing & royalties, custom silicon and connectivity products. As long as they are achieving the high margins they claim and the development demands can drop off a bit then there shouldn't be an issue with profitability.
The webcast was recorded and you can listen here: https://www.youtube.com/watch?v=_L4N0nzJ2wY
rz105, I agree with what you say. Hoping that from hereon there won't be any further financial disappointments. I did note however that they seem to expect to drawdown further on their cash balance this year, so year end 2024 may see a further lower cash balance although they also mentioned an expectation of higher revenue stream in the second half of this year. They expect 2024 H1 revenue to be less than 2023 H1 revenue as they continue to invest heavily in capital / R&D which is mainly focused on the 2nd generation opto-electronic connectivity product development for the tier 1 hyperscaler.
Tony did emphasize that the R&D is based on development of already proven technology as opposed to speculative research. He also mentioned that Alphawave are now number 4 in the world of silicon IP design (behind Arm, Synopsys and Cadence) in terms of revenue generation in that area.
Coincidentally, I've just found a recent link on this very subject:
https://semiwiki.com/semiconductor-services/ipnest/343875-semi-market-decreased-by-8-in-2023-when-design-ip-sales-grew-by-6/
No, they said there would be one today also.
Here’s the link:
Results Presentation and Webcast
A presentation for investors and analysts will be held today at 8.30am BST. The webcast will be accessible via:
https://awavesemi.zoom.us/s/84323327486?pwd=WFdWQzArdVBsN3JJcGlFbEM5WUo3Zz09
Passcode: 802056
If you go their website at 7am or just minutes after there will be an RNS here:
https://awavesemi.com/investors/regulatory-news/
In that RNS you will have the time of the session and there will be a link to the zoom session with a passcode.
If you not bothered to listen to the live session you will find a recording of the here about 30 minutes to an hour after the session has ended: https://awavesemi.com/financial-results/
Hope that helps.
It will probably be the usual crew: John (Chairman), Tony (CEO), Rahul (CFO). Jose will do the usual intro. They might bring in an non-exec like Jan.
Their current CFO is relatively new having joined last October.
https://tools.eurolandir.com/tools/Pressreleases/GetPressRelease/?ID=4405322&lang=en-GB&companycode=uk-alpw&v=
They also have a new Chief Revenue officer who only joined in February.
https://awavesemi.com/press-release/alphawave-semi-announces-appointment-of-charlie-roach-as-chief-revenue-officer/
The rest of the directors and management team can be read about here:
https://awavesemi.com/company/leadership/
And there's more detailed information on their backgrounds in FY2022 report from pages 72 to 77:
https://fr.zone-secure.net/28934/1843059/pdfs/Annual_Report_2022.pdf
It's fair to say that thus far they have over-promised and consistently under delivered where it matters (i.e on the financial reporting). Let's see if the two supposedly highly experienced recent recruits can turn it around. The preliminary update and investor call on Tuesday talked of taking a new bottoms up approach hence the downgrade in the outlook.
As soon as I saw the RNS this morning I made the reluctant decision to get out altogether. I've done so with a profit of around 18% over 2 years. Not exactly what I had in mind but better than coming out with less than what I put in.
It's likely now that I will wait until after the results before I make a decision to get back in or not regardless of whether price is higher or lower than now. I genuinely thought they would hit their guidance or exceed it. Further, I'm not thoroughly convinced of their reasoning for missing the guidance - accelerated exit from China? - that's a weak excuse. To be fair though it should be noted that missing the guidance by less than 10% isn't exactly disastrous but it does show they haven't got a proper handle on their finances yet.
Last month Alphawave did a presentation on advanced custom silicon for AI and Data Centres.
If you haven't already, it's well worth watching and there's a Q&A session at the end which always worthwhile listening to.
It's available to listen to here:
https://awavesemi.com/advanced-custom-silicon-for-ai-and-data-centers/
The presentation slides are available here:
https://awavesemi.com/wp-content/uploads/2024/03/Advanced-Custom-Silicon-for-AI-and-Data-Centers-2024.pdf
Alphawave are a significant key partner in the Arm Total Design ecosystem providing the connectivity IP.
The addressable market for custom silicon is currently about $9.2 billion this year and growing at a rate of 15% per year.
They issued an RNS back on 13th October 2022 which relates to the acquisition of Banias Labs and the non-binding framework agreement with the hyperscaler.
https://tools.eurolandir.com/tools/Pressreleases/GetPressRelease/?ID=4173559&lang=en-GB&companycode=uk-alpw&v=
"The Company expects the ramp of new Optical products to start in 2024 and profitability to improve from 2023 levels as products go into high-volume production and generate additional revenue. Alphawave anticipates additional revenue synergies given the complementary nature of the technology and its contribution to the product roadmap."
"The non-binding purchasing framework agreement defines a potential portfolio of optical DSPs, including coherent solutions based on Banias technology, that Alphawave intends to develop over the next several years. The Company anticipates ramping up sales to the leading hyperscaler over a multi-year period, with the potential for aggregate sales to exceed US$300 million, subject to the Company qualifying the products for use in the hyperscaler's data centers on a timely basis and ensuring sufficient silicon availability to meet its usage demands."
"In addition, the non-binding framework agreement contemplates the Company issuing future stock warrants of up to 2.5% of the issued share capital of the Company, with vesting based on purchases of Alphawave silicon products of up to US$700 million over the same period, and an exercise price based on the current trading price of the Company's shares, subject to downward adjustment if the Company's shares trade lower prior to issuance. The Company intends to solicit all required shareholder approvals, ahead of finalizing any warrant grants, which are not expected until at least 2024."
I've named the hyperscaler because it's been mentioned several times in posts by a number of well respected semiconductor bloggers including Semi Analysis and Fabricated Knowledge.
Comparing the technical products between these two companies is something I haven't as yet got fully to grips with.
It's true that they both offer high speed connectivity using the same protocols PCIe / CXL etc.
Astera Labs mainly offer retimers while Alphawave don't as far as I can tell.
When high speed data is transmitted over certain distances the signal quality deteriorates and retimers are devices that essentially clean up the data signal. It achieves this by amplification and then retransmitting a new copy of the signal. Market demand for retimers has been increasing as data quantity and speeds have been increasing over time.
Astera Labs have taken the approach of also offering a packaged software product called COSMOS which is designed to work with their full range of retimer products to provide a full suite of diagnostics and telemetry features enabling operators to optimize the data transfer performance.
Software is something that Alphawave don't do as far as I understand it. Yes they offer various PHY for SerDes, PCIe, UCIe die to die and HBM. They also offer the controllers and subsystems for these protocols including ethernet. The software for these however comes from elsewhere. This is more a case of allowing operators the flexibility to select the software product of their own choosing.
If I have understood correctly, it's possible that Alphawave's high end connectivity IP products achieve similar levels of high quality high speed data connectivity without the need for retimers and hence the possible reason they don't offer them as hinted in one of their press releases on their partnership with Nubis. Or, it is simply that it's a component that their customers incorporate from another of their partners elsewhere particularly in custom silicon designs.
While Astera Labs now have a ton of IPO money to invest into their business, their focus until now has only been in their retimer range with controllers and optimization software. No chiplets or custom silicon or opto-electronic investment on the horizon as yet. Astera Labs growth prospects stem from the partnerships they have with Amazon and Nvidia. At the current ridiculously high share price however and P/S ratio two to three times higher than their NASDAQ competitors I don't personally see them as an attractive prospect at this time.
Meanwhile, Alphawave's future is quite promising and currently generating more than three times the revenue of Astera Labs. They have an array of high profile partnerships including the multi hundred million dollar prospect with Amazon AWS which should hopefully see a range of opto-electronic products set for delivery this year. The current share price and P/S ratio being only half to a third of that of its NASDAQ competitors still makes it a far more compelling investment prospect.
"** Jefferies sees semiconductor equipment sector poised for solid Q1 performance with robust orders and optimism for
continued growth, despite slight demand softness in automotive and industrial markets.
Jefferies favors Dutch ASML ASML.AS , ASM ASMI.AS , and Swiss VAT VACN.S with "buy" recommendation for robust Q1
orders on strong DRAM demand and positive 2024 outlook, reinforcing sector confidence.
Also spotlights Alphawave AWE.L and Aixtron AIXGn.DE as top small-midcap choices with "buy", expecting strong Q1
order intake and resilience in tech advancements. **"
From Thomson Reuters.
Yes you could be right there. The X330 has CoWoS packaging (as will the X340) but the mention of that in relation to the Nvidia chips could just mean like similar to how they have CoWoS on the Nvidia chips. It's vaguely worded as well as being a translation from Korean.
Curious as to why neither AWE nor Sapeon have issued a press release regarding the deal that was done in February. The fact that it is on the Korean ELEC website presumably suggests it's not confidential? Perhaps at $44 million value it's small fry news... Still every little bit counts. Interesting mention of the X330 being used in Nvidia's A100 & H100 chips. Could the X430 wind up being used on future Nvidia chips?
Hi Awave,
At this present stage we only know what you already know and that is they plan to launch this year. I'm hoping that they will have something to say or reveal at the next trading update later this month as to the progress they have made.
As you probably know, the Banias labs acquisition came with a non-binding agreement with a hyperscaler whereby if these Opto-electric products pass all the necessary qualification trials (which by now might already be largely complete), these products should in the production phase for delivery sometime during this year. In the Banias Labs acquisition announcement back in October 2022, the expectation seems to be that they should be able to sell a minimum of $300m as the sales ramp up over the next several years. It also mentioned that AWE would issue warrants to to the hyperscaler up to 2.5% of the issued share capital up to a maximum of $700m in sales. It is however a non-binding agreement and so if these products don't deliver on expectations there is the risk of pull out.
Investor relations told me that while these products are initially intended for the hyperscaler with whom they have the agreement they are standard products that will also be available to other hyperscalers and data centre operators. The company have also engaged with key module makers which is creating opportunities with other customers in the opto-electric segment. The product range will include not only DSP transceivers of both PAM4 and coherent but also Active Electrical Cables and Active Optical Cables as well as other types of opto-electric products. Incidentally, these products are expected to generate high gross margins in the range of 60 to 70%.
Interestingly, AWE issued a press release on Friday regarding a collaboration with Innolight where the intention is to implement their PCIe technology IP into Innolight's optic products. Whilst Innolight are a Chinese producer, they are nevertheless the market leading producer of DSP transceivers globally. The competition is stiff though as Innolight announced a similar collaboration with Marvell at the same time.
Another correction. Sorry!
Astera Labs total shares outstanding following today's IPO will be 135.7 million which at $61 per share gives a market value of nearly $8.3 billion. And thus based on $116 million revenue gives us a P/S of 71.4. Alphawave's equivalent price on that ratio would be £26. So still absolutely bonkers. I should not be surprise to see ALAB price take a tumble in the coming days because if not, it would mean the world doesn't want to make sense to me any more...
CORRECTION.
I made the mistake of assuming that the 19.8 million shares would be the total shares in issue. But in fact, having checked the SEC website, it is more like 155.5 million shares. Assuming I've got that right, this therefore puts a market cap value on Astera Labs at a crazy $9.3 billion. It means a P/S ratio of more than 80! Which is completely bonkers because if AWE was priced on that ratio the share price would be something like £29!
With 19.8 million shares at current price around $61 puts the market value at $1.2 billion. As Jupiter31 mentioned they generated revenue of $115 million for 2023. (Not sure which month their financial year ends). This gives them a P/S ratio of 10.6
Alphawave are currently valued at almost $1.6 billion and in next month's results will likely report revenue of somewhere between $340 million and $360 million. At midpoint of this revenue gives a P/S ratio of 4.6.
If Alphawave were to be priced at P/S of 10.6 we'd be at £3.86.
Astera Labs appear to have a slight different approach with their connectivity solutions. I'm no expert on this, but for example, one of their key products are high speed retimers - these devices modulate the signals by restoring and reshaping them as they pass through the link. Essentially they are mitigating the signal degradation that you get due to attenuation, jitter and noise. Whereas Alphawave have products that can achieve the same (or better?) mitigation of the signal without having to use a retimer such as in this press release:
https://awavesemi.com/press-release/nubis-communications-and-alphawave-semi-showcase-first-demonstration-of-optical-pci-express-6-0-technology/
It's worth remembering that Astera Labs have been around awhile and are not suddenly new competition for Alphawave to contend with. One aspect of Astera labs that is attractive is that they are in partnership with Nvidia for some of their products. It will be interesting to monitor them in the coming months and years to see how they do.
It will also be interesting to find out what Tony Pialis might have to say about them at the next investor call on results day sometime next month, as I'm sure one or two of the analysts on the call will bring up the question.
Does anyone have an idea please of why Alphawave shares have dropped over 10% this past week?
It's had a bull run with approximately 72% increase since January. So it's taking a breather along with all the other "AI" stocks which have declined by similar amounts in recent days. The drops are more notable from around the time of the US markets opening time. Even Nvidia has taken a drop. Nothing to do with the fundamentals of company.
Credo has more than 2x the market cap of AWE despite only generating half the revenue. They compete with AWE in the connectivity space but have more focus on AECs (Active Electrical Cables) which amounts to approximately half their revenue. Like AWE, Credo are ploughing back a lot of cash into R&D and are not net profit making. Credo may appear to have a healthy cash balance sheet but only due to having done a recent cash placing. They are in partnership with Effect Photonics for the development of Coherent DSPs but I believe are behind the curve compared with AWE in bringing these to market.
Marvell has a market cap more than 37x greater than AWE’s. They are AWE’s main competitor in terms of data centre connectivity products. Last Thursday was Marvell’s Q4 results which gave a very upbeat performance in the data centre sector. Unfortunately, this was offset by a shockingly downbeat performance in their carrier infrastructure and consumer sectors neither of which AWE compete in so far as I’m aware. That probably explains the market’s reaction to the Q4 results that brought about a more than 11% fall in share price on Friday. Similarly to both Credo and AWE, Marvell are also not currently net profit making. Again, mainly because a large amount of cash is being ploughed back into R&D and product development.
While the two above-mentioned NASDAQ peers have in my view excellent growth prospects, I think AWE’s growth prospects are at least equally as good. As mentioned in my previous post they should be due at some point later this year to announce their first revenues coming from their Opto-electronics range which will include coherent DSPs, AECs and AOCs. While Marvell appear to be slightly ahead of the curve having already brought coherent 800 Gbps DSPs to the market and are introducing 1.6T Gbps DSPs later this year, I suspect that AWE are not too far behind. The revenue stream expected from this is expected to ramp up quite quickly in the following years.
According to a fairly recent Whitepaper from IPnest, in what’s called the High-End interface silicon IP segment, AWE already command a quarter of that particular addressable market. Back in 2021 this segment was valued at about $370 million and is thought will have grown to around $1.4 billion this year. Assuming no increase in market share, this translates to about $350 million of revenue from that segment alone, add to that, revenue from custom silicon, chiplets and the new opto-electronics range, we shouldn’t be surprised if the company up their 2024 year end guidance by a significant amount, hopefully in excess of $400 million. By 2026, the high-end interface IP segment market is forecast to exceed $2 billion. The whitepaper predicts the company’s revenue will be between $500 million and $800 million. Total product revenue is expected to exceed $1 billion. All the above products have high profit margins.
Just over 12 months ago on 24th Feb, Alphawave closed at an all-time low of 89.3p. From that baseline, the share price is up 110.3%. If we take the price of its Peers on that same date of 24th Feb last year and compare with their respective closing price on last Friday, Rambus is up 134.41%, Broadcom is up 126.57%, Credo is up 111.2% and Marvell is up 72%.
So, one can argue that Alphawave has performed equally as well as its peers over this time frame.
Another interesting comparison I like to make is the price to sales ratio (P/S).
If we take Alphawave's 2023 midpoint guidance revenue of $350 million, last price of £1.878, and 727 million shares in issue. We get a P/S ratio of 5.02 which is significantly lower than its peers.
Based on the absolute latest data from the peers last reports and £to$ conversion of 1.287:
Rambus P/S = 15.25. If AWE were trading on this P/S, its share price would be £5.71
Broadcom P/S = 16.94. If AWE were trading on this P/S, its share price would be £6.34
Credo P/S = 19.7, If AWE were trading on this P/S, its share price would be £7.37
And Marvell P/S = 11.86. If AWE were trading on this P/S, its share price would be £4.44.
Taking the average P/S ratio of these four peers, AWE share price would be £5.97.
Even though AWE has been on a price surge since January 17th and risen 72% since and no doubt various technical indicators are currently screaming "overbought", but regardless of whether a price correction is due or not, at this current price of £1.88, it remains significantly cheap compared to its peers. I'm not a technical analysis fan, but I've seen that the future growth indicator on this stock places it at the top end of the growth shares table.
At some point this year, the company are expected to announce its first revenues from its opto-electronics range including coherent DSPs for its hyperscaler with whom it has a non-binding agreement for at least $300 million worth of product sales in the pipeline. All things data centre connectivity product wise have fairly high CAGR expectations over the coming years but I believe the coherent opto-electronic aspect for intra data centre connectivity applications especially so.