And then October and Q4 rolled around. Stock markets are still going strong. Nvidia is still at $3trn. And OpenAi is looking to raise $6.5bn at a $150bn valuation. And, oh – by the way – does the SaaS model have a future in the new AI paradigm?

Read on for an analysis of what’s going on in the world of tech and venture as we open the last quarter of 2024.

Nvidia holding steady at $3trn

The S&P500 is up just shy of 2% in September, with Nasdaq 100 up 3%. Both are hovering close to all-time highs.

US inflation appears to be coming under control. We got the 50bps rate cut from the Federal Reserve, with another 50bps forecasted before year-end. Currently, it looks like rates will go to just shy of 2.9% over the next two years. This is much higher than the 0% we had during the ZIRP era but much lower than the +5% we’ve had until recently.

Nvidia is still flying high.

Nvidia is also up ~3% in September. The stock has been in a holding pattern since June.

A few months ago I discussed the bull and bear case for Nvidia at $3trn. While the stock still looks highly valued at a P/E multiple of 57x, many investors continue to see significant upside in the stock.

What’s underpinning the continued optimism?

In September, we had signals from industry leaders as to why there continues to be so much excitement behind the stock.

Marc Zuckerberg from Meta outlined what’s so special about the new generative models. Whereas past machine learning models have hit ceilings in terms of performance, we have not yet seen such ceilings for the new transformer model (“transformers” are core to the technical architecture that underpins Large Language Models like the model behind ChatGPT).

More data and more CPU have kept leading to better performance.

The giants are battling it out

x.ai (Elon Musk’s AI firm) has just brought a new 100,000 CPU cluster online – the first in the world. And he has started talking about 200,000 clusters.

The cost of the new Nvidia Blackwell GPU could be as much as $60,000 – $70,000 per chip.

In other words, one 200,000 CPU cluster will cost $13bn in chips alone.

And the whole world is racing to get these chips.

Microsoft and Alphabet have essentially said they are going to keep spending on these. Nobody wants to be left behind. Sergey Brin from Google is now back in the lab full-time. He sees AI as the biggest sea-change in technology, at least since the Internet.

In other words – looks like the industry will keep spending on Nvidia chips for the foreseeable future. So Nvidia could be a good buy, even at $3trn.

What’s the case for OpenAI at $150bn?

Sam Altman (OpenAI CEO) has been in the news a lot lately. He is currently working to finalise a new investment round in OpenAI that would raise another $6.5bn from investors at a $150bn valuation (alongside $5bn in debt), up from $86bn valuation earlier this year.

The minimum commitment in the round is set to be $250m, so this one possibly won’t fit right into a standard EIS round(!)

From not-for-profit to $10bn’s worth of shares

As part of the round, Open AI would convert from a not-for-profit to a B Corp, and, among other things, remove the previous 100x profit cap for investors.

As part of the plan, Sam Altman would go from being the altruistic steward of said not-for-profit to having a 7% stake in the $150bn company. There is a meaningful difference between having shares worth zero and having shares worth $10bn. Good for him if he can pull it off.

Bull and bear case

But it raises the bigger question – what is the bear and bull case for OpenAI at $150bn?

Bring out the bulls:

1. Incredible traction: OpenAi’s ChatGPT remains the fastest growing B2B software product in history, growing from launch to $1bn in revenue in an estimated 8 months (across all software, the crown is held by Grand Theft Auto V that scaled to $1bn in just three days). Current forecasts say that OpenAI will be at $4-6bn of run-rate revenue by the end of 2024 and that it could be at $6-10bn by the end of 2025. $150bn valuation at $10bn of run rate revenue would imply 15x forward revenue multiple. Not unreasonable if the growth continues.
2. Technology leadership: OpenAI was first to market with a commercial Large Language Model, and the company has continue to innovate in everything from voice to image generation and video. The company’s continued tech leadership means that the company has developed the biggest mind-share in the developer ecosystem, with others building on top of their platform. Over time this could develop into a moat – especially once they add data storage to their offering (one of the issues with current models is that they don’t offer a straightforward way to store data. OpenAI is working to solve this).
3. Industry leadership in a massive market: As anyone who has been close to the AI transformation will tell you, AI has the potential to be the biggest transformational technology of our lifetime. Sergey Brin (Google co-founder) is now back in the lab full time working on AI, because he sees the opportunity as so significant. PwC estimated that the annual opportunity from generative AI alone is $15trn within a decade. Whether or not this is the right number, the consensus from those who are close to the transformation is that this is going to be big. If you think AI is going to be even bigger than search, the leader could be immensely valuable. And OpenAI is currently the leader of this industry.

Line up the bears:

1. Intense competition and risk of commoditisation. Open source LLM’s like Meta’s Llama are already incredibly good. And they have become relatively easy to embed into other application frameworks. There is a real risk that LLMs could become commodities which could destroy OpenAI’s business model.
2. Financial disparity with competitors: As outlined above, the cost of GPUs to train models has become very high. And due to the scalability of the models, there is a possibility that the winner will be the firm that can fund the biggest training clusters. This is unlikely to be OpenAI. Meta generates $40-$50bn of free cash flow per year. Alphabet generates $60bn. Microsoft $75bn. It will be challenging for OpenAI to raise sufficient capital to outspend the behemoths over time.
3. Organisational challenges: OpenAI has had excessive leadership turnover this year. So far the company has lost Mira Murati (CTO), Bob McGrew (Chief Research Officer), Barret Zoph (VP Research), Ilya Sutskever (Chief Scientist) and Andrej Karpathy (Co-founder). Can the company keep it’s pace while losing so many leaders? And are the departures a symptom of underlying issues in the business?

The cult of Sam

It’s clear that OpenAI is one of the defining companies of our time. And as the company’s notoriety has grown, Sam Altman’s image has transformed from a man on a scientific mission to bring about Artificial General Intelligence for the benefit of all, to a hardnose tech tycoon who isn’t afraid to leave collateral damage behind as he drives to make OpenAI the leader of the new AI landscape.

In some ways, Sam Altman is starting to remind me of Elon Musk. Elon also had plenty of setbacks and staff turnover at his companies, but he keeps defying his doubters to hit new milestones in his companies. And just as it is impossible to write off Elon Musk (whether you like him or not), it feels difficult to write off Sam Altman and OpenAI. There is plenty of risk, but the upside is monumental, and the company keeps executing.

What can we learn from the cloud computing wave?

A few more notes to draw parallels between the new AI paradigm and the cloud paradigm that came before.

When the cloud took off with the launch of AWS, many people initially thought cloud computing would commoditise quickly. Intel-powered cloud servers were already on their way to complete commoditisation. And Linux was already a free open-source operating system. But today it is Intel that has been commoditised. And AWS offers more than 200 district products, with Amazon having created a complex ecosystem with more than $100bn in annual revenue.

Commodity? Maybe. But ask anyone who is trying to compete with AWS, Google and Microsoft in cloud computing. Those three companies have made formidable cloud businesses, and others have fallen by the wayside.

And while OpenAI at $150bn might not be a slam dunk investment case, I wouldn’t bet against Sam either – especially if he can continue OpenAI’s impressive revenue growth.

One to watch closely over the coming quarter.

Does the software industry have a future in the age of AI?

The Future of SaaS in the Age of AI

Outside the tech industry, many people still discuss whether AI will have a big impact on the world. For those of us who are in the tech world, that question seems settled. I use LLM’s many times every day. They have replaced 80% of my Google searches. They help with ad hoc analysis, code development, process automation, and so much more. As I wrote a few days after the first release of ChatGPT, the arrival of the LLM’s changed the world forever. And that is indeed the trajectory we continue to see.

Now, in some ways, we are still only scratching the surface. But a big question inside the tech industry is whether AI and large language models (LLMs) will make the traditional SaaS industry obsolete. Some argue that as AI becomes more advanced, businesses will rely on AI agents and open-source databases instead of specialised software. Is SaaS really on the way out?

Is SaaS Disappearing? Let’s look at the arguments for

First off, AI agents are becoming more capable. They can automate complex tasks that used to require specialised software. For example, fine-tuned LLMs can now handle customer support inquiries. This could make dedicated support software unnecessary.

Another aspect is that AI models can analyse vast datasets in real-time. This could make traditional enterprise systems like Salesforce or Oracle less relevant, as businesses could use AI to interpret data directly without needing those systems.

Also, as AI makes businesses much more efficient, we may not need as big companies anymore. And if we don’t need massive organisations, we might not need all the software that goes with those organisations either.

Lastly, open-source databases combined with AI could cut costs significantly. This challenges the SaaS subscription model, as businesses might prefer free or cheaper alternatives.

Why SaaS Will Thrive

Despite these points, I believe the SaaS industry isn’t going anywhere. Here’s why:

Specialised Solutions and Compliance

SaaS providers offer software tailored to specific industries with unique challenges. Take healthcare, for example. SaaS platforms manage patient data while complying with strict regulations like HIPAA. Generic AI models can’t easily replicate this level of specialisation and compliance.

Reliability and Support

SaaS providers offer guaranteed uptime and dedicated customer support. Businesses depend on these assurances for uninterrupted operations. AI models lack the accountability and support infrastructure that SaaS companies provide. The open-source world has plenty of parallels. Linux is free. But enterprise customers still buy Linux support from IBM/Red Hat for billions every year.

Limitations of AI Models

While AI is powerful, LLMs can produce inaccurate or unpredictable results (“hallucinations”). In critical applications where accuracy is non-negotiable—like financial reporting or compliance management—these risks are unacceptable. AI models also often lack explainability, making it hard to justify decisions in regulated environments.

Another related aspect is that while many new areas can be automated using next-generation AI models, this won’t be done by generalised LLMs. Rather, it will take specialised models trained on proprietary data. Far from obviating specialised SaaS, this opens up whole new SaaS market opportunities, as new markets will grow with the new capabilities.

Cost and Resource Efficiency

Developing and maintaining AI solutions requires significant investment in expertise and infrastructure. SaaS offers a cost-effective alternative by providing ready-to-use solutions. That equation is not changing. It’s still more efficient for specialist SaaS vendors to develop a solution and sell it to many customers, than for each customer to handcraft their own systems.

Conclusion

Rather than being replaced, SaaS is integrating new AI capabilities to deliver even stronger value propositions. This synergy offers the best of both worlds—advanced AI capabilities within a secure, compliant, and supported framework.

As a result, the SaaS industry will continue to thrive despite the rise of AI. While AI will undoubtedly transform the software landscape, SaaS offers specialized, reliable, and compliant solutions that AI models alone can’t fully provide. The future lies in the integration of AI within SaaS platforms, ensuring businesses reap the benefits of AI without sacrificing the support and security that SaaS provides.

So, while AI will play a significant role in shaping the future, SaaS isn’t disappearing. There is no doubt that some SaaS players will come under pressure, and some even perish. But as AI allows us to tackle ever more use cases, it will lead to a growing software market. And this will lead to growth for the overall industry.