Google Cloud Next London 23: AI at the heart of the cloud

PAC recently had the opportunity to attend the Google Cloud Next London 23 event. The event allowed Google to emphasise what they consider to be their differentiators in the highly competitive cloud hyperscaler market segment. It was an interesting event for PAC because the overwhelming focus of the messaging throughout the day was data and AI-driven capabilities across the breadth and depth of their solutions.

For an event titled “Google Cloud Next London 23”, it was surprising to PAC how little the sessions provided to analysts discussed the cloud directly. PAC’s interpretation of the messaging provided is that the Google Cloud Platform (GCP) differentiator is the laser-like focus on providing organisations with a wide range of capabilities to adopt and scale AI from across a range of model types, including generative AI (GenAI). At the analyst-specific sessions, a question referred to understanding how Google competes with the two biggest cloud hyperscalers in the UK and Europe. Now, whilst that is an understandable perspective, and our market data reflects this too, what was interesting to PAC, as the sessions throughout the day occurred, is that Google could potentially redefine the hyperscaler market dynamics.

Is the future of hyperscalers raw compute or AI scaling?

At the event, one of Google’s European clients said they consider Google Cloud the best hyperscaler to support data-related deliverables within their business. PAC considers this a prime example of many organisations’ transitional journey. The age of digital transformation has seen organisations transition to a wide range of cloud operating models from dedicated software-as-a-service (SaaS) companies to compute hyperscalers. From PAC’s perspective, organisations have typically retained the legacy system approach to operational IT management during this period but transposed into compute hyperscaler capabilities. This has driven a race between those companies wanting to be considered hyperscalers to offer an ever-expanding and often overwhelming range of cloud-based compute services. The typical enterprise pattern is a transition to the cloud, then a process of re-architecting and re-engineering to drive further value from a cloud-based IT operating model. This, PAC considers, has led the industry to a point where it is hard to differentiate hyperscalers between their “all you can eat” messaging.

What PAC found refreshing at the Google event was their laser-like focus on data and AI services, delivered through one of the best cloud hyperscale compute platforms, as the differentiated messaging for organisations looking to consider who is best to engage with. It was discussed at the event that there is still a wide range of maturity regarding the cloud across Europe, from fast followers to laggards. However, PAC considers this expected, and the transition to the cloud is no different from any other positively disruptive technology over the last forty years. It is natural that despite a widespread acceptance and understanding of a technology like the cloud, specific industries, countries, and organisations will be limited by their individual comfort in adopting a technology. This, though, from PACs perspective, should not be considered a negative issue relating to Google because they can educate their prospective client base directly and indirectly through their partner eco-system. Ultimately, it is not Google’s fault that organisations in 2023 can’t determine the value of the cloud for their organisation and treat it as something to be suspicious of despite its globally ubiquitous usage.

Over the past thirty years, PAC has witnessed the three distinct ages of the internet, mobile, and data. All of these shifts led to the volume, variety, and velocity of data becoming both a transformative differentiator and also, for many, an overwhelming operational burden to process. While GenAI has driven a broad understanding of the sophistication of AI to mainstream public awareness, AI’s role in optimising and scaling the value that can be derived from data has been clear and evident over the past decade. It is clear to PAC that the critical value a hyperscaler provides organisations in 2023 and beyond is not purely compute but rather the ability to operate and scale AI to meet both the bespoke and productised needs of an organisation atop a hyperscaler like GCP.

This was interestingly reflected at the event by the following list of questions Google presented to attendees regarding what prospective and existing clients are asking them:

• Are we partnering with the right technology providers who will position us for success today and in the future?
• How do we make Al accessible to everyone in our organisation?
• Do we have the best technology infrastructure to build scalable systems that work better together through an open ecosystem?
• Are we using the best tools to boost the productivity of our employees to leverage Al-powered tools and workflows?
• Are we confident that our data, which is our ultimate differentiator, is protected and secure as we leverage new Al tools?

Establishing a foundational approach to AI

Due to the events late last year relating to GenAI, 2023 is considered by PAC to be the year that the broad awareness of the sophistication of AI became tangible for people in their personal and professional lives. However, many of those same people are likely unaware that the use of many forms of AI is not a phenomenon of 2023 and instead has been permeating our personal and professional lives through an ever-increasing range of ways we engage with digital devices and services over the past decade.

It is clear to PAC from the Google event that the company knows this all too well and represents itself accordingly, given the rapid onset of increased attention regarding AI. Historically, PAC has seen with other innovative technologies that they rapidly become productised across various solutions and services. However, the GCP approach stands out to PAC for organisations requiring a specific and sometimes bespoke capability for AI. The Google Vertex AI product provides all the benefits needed for an organisation to drive AI at scale to meet their needs without incurring traditional levels of operational complexity that have occurred over many prior decades with bespoke developments. A particular standout to PAC is Google’s approach to providing organisations access to 100+ foundation AI models spanning first-party, open-source, and third-party channels surfaced through what Google calls its ‘Model Garden’. This allows organisations to pick whichever model, from whatever source, that meets their needs and switch to other similar models if a foundation model of the same AI type becomes more sophisticated and thus more relevant to an organisation’s needs. One of Google’s clients stated that this flexibility to pick and change foundational models from across a wide range of AI types and within specific AI segments was a competitive differentiator as to why they considered Google the most suitable hyperscaler for AI adoption and scaling.

At this point, PAC concurs with this sentiment because the ‘Model Garden’ allows organisations to leverage ever more sophisticated foundation AI models without the operational cost and complexity of managing this. Google acts as the trusted partner, leaving the organisation to benefit from leveraging the best AI models as they become available. PAC considers minimising the complexity of AI adoption whilst maximising the scale of operationally using it to support business processes and workflows as the core value battleground for hyperscale providers going forward. To that end, Google presented a compelling argument as to this.