The AI Pivot: Why immersive tech companies need to tread carefully
The rise of AI is positive, yet there are large risks ahead.
Firstly, I get it. We are now a year in from the launch of ChatGPT, and I don’t need to write yet more articles on why AI will be impactful, especially for our industry. You know it, I know it, and you don’t need me to convince you.
But the heart of this piece is different; I think we are pivoting towards AI too hard, and it will undermine us in the long-run.
I understand getting caught in the wind of positive change – the same happened with metaverse-related companies in 2021. But I worry that a hard change will ultimately undermine many companies. I will touch on why I think this is, and what I recommend for companies to stay the course in 2024 and beyond.
Integrating AI: Immersive tech companies are adopting AI to augment their services, especially during slower growth periods, as seen with companies like Meta and VirtualSpeech. That is fine in itself, but there are traps ahead.
The GPU shortage: The scarcity of GPUs poses a challenge to AI development and the creation of virtual worlds, potentially leading to high costs and limiting smaller companies' access to the tech.
Careful scalability: While embracing AI, companies should ensure that their growth is manageable, and that AI integration does not lead to unsustainable costs or overshadow their foundational business model.
Balancing with core values: Don’t prioritise AI at the expense of their core mission, risking the loss of their unique value proposition in the market.
The AI Pivot
We’ve already seen a few companies that have already pivoted towards AI and received a satisfying stock boost in the process. Meta outlined a series of announcements relating to its AI models and integration with its hardware, which boosted its stocks by 11% in May (alongside positive revenue). The same goes for Google in August, when it announced a slate of AI offerings for Cloud and Workspace platforms.
AI is helpful in numerous ways, so it’s no surprise that immersive technology companies are integrating its capabilities into their services. Plus, the stock climate coincides with a quieter period on immersive technologies, where growth is less exponential. Some companies that have integrated AI over the last few months:
Meta, integrating its AI services into its new Ray-Ban specs. The company is also exploring how AI can generate 3D models.
VirtualSpeech added AI capabilities for soft skills training, such as interview experiences.
Simulacra, who released AI tools to generate imagery for architecture and clothing.
ENGAGE, who is using ChatGPT to create an AI assistant called Athena.
(I haven’t mentioned Nvidia or its Omniverse platform, as it feels as redundant as saying that clouds appear during a thunderstorm).
I have no doubt that AI will be a huge feature in metaverse-related companies in the future. It provides a frictionless layer to designing virtual worlds, and a seamless way to converse with environments and experiences. Integrating AI will become the norm. But I worry about an over-reliance on its capabilities to help boost visibility and viability, for two reasons: The GPU Throttle, and losing a company’s heart.
The GPU Throttle
I will write a full piece on this soon, but the lack of GPUs is a seismic limiting factor for nearly all companies working on virtual worlds or AI. There is a direct lack of GPUs that can handle the workloads required to sustain virtual worlds, or to run AI applications. The lack of GPUs is so stark that Microsoft is working with Oracle to increase capacity for its Bing-related searches.
Nvidia is working hard to meet the demand, tripling the production of its hotly-contested chips next year, and it does not include how competitors like Intel may step in with their own offerings. But until then, many smaller companies are priced out by its high costs. When OpenAI announced that its API costs had been slashed, it garnered a positive cheer during its developer conference. High costs are a significant impediment to business adoption. And I haven’t even mentioned the severe lack of relevant talent like data scientists.
My worry is that immersive companies should be careful to pivot so hard into AI when costs could become untenable in the long-run. Running multiple enquiries is difficult to do at scale, as processing power is so high and costs can be difficult to manage. While companies are smaller, the costs justify a well-timed press release or positive announcement; but at scale, only the larger players have the finances and support to handle the load. It’s like promising the world’s best doughnuts, but running the doughnut factory kills the business when too many people notice.
Losing your heart
The other reason I worry is that companies may lose the heart of their value proposition. I’ve noticed companies in my inbox pump AI stories like there is no tomorrow, almost dwarfing its heart with its new vision. The stories are positive, but it is a path that paves its way towards losing the core of the business. I haven’t yet seen any immersive tech companies who self-sabotaged their way to oblivion publically. But outside of it, I have seen sites like Gamurs Group almost self-implode when it proposed firing its team to leverage ChatGPT and a copy editor. It’s a fate which I hope our industry avoids.
Many companies are in the game because they wish to provide compelling and immersive virtual worlds or tools, which heighten the human experience. AI plays a role in that, yes; but it is a tool towards the whole story. So my messaging advice is:
Celebrate the benefits of AI and how it works with your products and services;
Keep to your core value proposition, and make it crystal clear that it is part of your story, not your new one;
Keep a level head with both users and investors, ensuring that scalability is sensible.
The advice is not revolutionary or surprising. But it is important to point out the obvious as we continue our journey into the new world of AI. Let’s keep a cool head as we focus on what we are good at.
Note: The Immersive Wire is run by Tom Ffiske, who also works within Accenture’s metaverse group. The contents of the newsletter should not be regarded as Accenture’s views. Images were created with DALL·E 3.