Skip to content.

Tech trends in 2021

Perspective, by AlbionVC


As we look ahead to the new year, here are our technology predictions for 2021. 

A flood of ‘advanced therapies’ to be marketed for the first time 

A flood of so-called ‘advanced therapies’ (including cell and gene therapies) will complete clinical trials in 2021 and be marketed for the first time (just a handful are currently available in a small set of disease indications). Many of these products will carry enormous ‘sticker’ prices (in the hundreds-of-thousands or indeed millions per treatment) but the benefit of these curative therapies is many multiples of the cost, both for inherited diseases as well as in cancer treatment, with continued strong reimbursement trends to be expected. A second generation of these approaches – with improved safety and efficacy – is slowing working its way through the development process; gene editing and modular cell therapies are already getting into the first stages of clinical testing. Vaccines will of course be a big focus, beyond just Covid-19, as several technologies mature and the path to market becomes accelerated. Bi-specific and tri-specific antibody therapeutic approaches are expected to continue to proliferate, especially in oncology and in autoimmune disease.

AI and Data Stack: ‘AI unleashed’

In a world that is increasingly becoming AI centric and where data and tools are more widely available, we are excited by the switch of companies from traditional software to more advanced data platforms. We believe that with that, comes a need for an advanced unified stack to support and optimize ML initiatives in Enterprises in order to get the promised ROI: from Data Ingestion, Data Preparation, Data Storage, Control Access, Building Insight and better models, Model Management (Deployment and Monitoring), and Explain-ability. We have already placed bets in some of these areas, through exciting companies such as:  HumanLoop (Active Learning); Seldon (Predictive Model Management Infrastructure); Imandra, (Automated Reasoning Platform) and continue to be on the look for amazing founders who are trying to revolutionize that space. 

B2B marketplaces are coming

Nearly all marketplaces successes so far have been in B2C while the vast majority of global B2B spend is still offline. B2B marketplaces are therefore in their infancy but we see more and more emerging across many sectors to digitise manual and analogue workflows. B2B marketplace sales are expected to grow by 6x in 5 years to reach $3.6tn in 2024 and surpass B2C marketplace sales by then. This is driven by a new generation of business owners and employees who now expect to buy online, more mature payments / API infrastructure and software development. These B2B marketplaces will be typically more SaaS-enabled than B2C marketplaces and will create tremendous value for the supply and demand sides by building software tools for them that simplify and automate time-consuming procurement processes. We believe 2021 will see more B2B marketplaces emerge at the start of what could be the decade of B2B marketplaces.

NLP use cases to increase rapidly 

Natural Language Processing (NLP) has seen a number of very exciting developments on the back of huge investment in the past few years, ever increasing processing power and advances in machine learning techniques.  NLP splits broadly into Natural Language Understanding, which translates language to data and enables it to be analysed, and Natural Language Generation which creates the right language based on data.  As these two technologies are developing, they are now at a point where they are accurate enough to open many use cases and improve the effectiveness of existing ones.  For example, companies such as Black Swan can analyse vast amounts of unstructured comments from social media channels to identify new trends in consumer behaviour or predict emerging trends, so that their customers can develop or reposition products to meet new demand.  Similar techniques can be used to manage credit risk in supply chains or develop investment trading strategies.  Evermore sophisticated chatbots are transforming the customer services sector across all industries, saving cost and improving on the customer experience. 

On the Natural Language Generation side, companies such as Phrasee automatically create marketing copy (email subject lines, email text, website landing pages, push notifications, Facebook ads language) which is optimised for specific situations or campaigns and performs significantly better than marketing copy created by humans.  This translates to millions of incremental revenue for its customers.  Other use cases of NLG include creating media articles from sets of data (e.g. sports events reporting) or summary reports from BI systems. 

As NLP continues to develop, it will become the prevailing interface between humans and machines.  We are already seeing this, to an extent, with Siri and Alexa but more complex use cases are being developed and adopted.  New machine learning techniques are allowing NLP models to be trained on smaller data sets, which opens up the potential for many more use cases.

Novel telehealth approaches for chronic conditions as healthcare providers embrace digital

New types of “digital-first” care models have recently started emerging focused on chronic health conditions such as metabolic diseases, heart disease and mental health. These models are combining software, hardware and specialist therapists to deliver an integrated, data-driven, patient centric, remote service that leads to improved outcomes at significantly lower cost. Covid-19 has significantly accelerated this trend as remote care went from the fringes to the mainstream in 2020. In 2021, we will see a broadening of these models into other disease indications, for example respiratory, gastro-intestinal and others.

Quantum computing turning up the volume

We’ll see more qubits on chips particularly from the superconducting computer superpowers, IBM and Google, but that won’t necessarily translate to huge leaps forward in performance as error rates will still be high. In response we’ll see ‘quantum volume’ enter the mainstream lexicon – this is an IBM created metric which accounts for both the number of qubits and how many steps can be performed by those qubits before they decohere (circuit depth). Despite the limitations posed by low circuit depth the smartest quantum theorists and software developers will build on Google’s demonstration of supremacy and start to show what quantum computers can do on a wider set of problems.

Remote work remains key focus for enterprise security professionals

Remote work is here to stay. Given this, the old paradigm of perimeter-based security has been turned inside out. Enterprise security risk now sits with employees. 

2020 already revealed a spike in adversarial activity targeting individuals in order to penetrate the enterprise. We expect this trend will continue through 2021 and beyond. We therefore anticipate enterprises will prioritise spend on technologies that support zero-trust models i.e. those that remove any notion of implicit trust based upon an individual (how we have traditionally accessed corporate networks), and require authentication and authorisation before establishing any remote session. 

In practice, this requires an entire re-architecting of enterprise security to involve new technologies such as federated identity management, biometrics, identity as a service and risk-based and step-up authentication. Ultimately, identity is the new perimeter.

The “win win” FinTech model

FinTech today exists predominantly as digital representations of Financial Services that have been around for hundreds of years. While admirable, and no doubt consumers and businesses are benefiting from this, revolutionary innovation it is not. Why is that? Financial Services’ incumbents and in many cases, insurgents still have a “product” mindset. Consumers and businesses alike just do not care about financial products they care about financial outcomes.

Enter embedded FinTech. Due to increasing confidence in managing financial matters digitally and advances in technology we are on the cusp of the nirvana – engaging with Financial Services natively through touchpoints in our everyday lives. Advances in identity and verification, cloud and mobile infrastructure, API connectivity and interoperability, and data enrichment are enabling technologies to increasingly act harmoniously in concerts to deliver seamless end-to-end user experiences. Payment infrastructure to enable micro usage, predictive and prescriptive financing solutions and real time embedded insurance are nearing production grade and will advance in 2021 making Embedded FinTech an exciting investment arena.

Why is this a win win? The end user benefits are obvious, but all value chain participants must benefit to make this a reality, be it at the manufacturing or distribution stages. For those that operate close the end user the benefit will be transformative: Embedded FinTech creates the possibility of additional monetisation which can significantly alter the unit economics of the original value proposition making it inherently more profitable. Previously uneconomical market segments can be entered increasing absolute dollar profit, or relative margins can be expanded from an existing core segment thanks to the near zero marginal cost of the additional monetisation streams. Manufacturers benefit from dramatically lower customer acquisitions costs from the hub and spoke nature of their now distribution customers. So where is the magic? This is the power of truly connected technologies, increasing the amount of friction being removed from the value chain and that means lower costs for all.

With all the battlefields up for grabs it will be interesting and exciting to see who can capture the value – if you think you are we would love to meet you!

Related content

WeGift closes £26m Series B funding as demand surges

Read the article >

Welcome Kibriya and Valerie to the B2B software investment team

Read the article >

Why we invested in Toqio?

Read the article >